Description of Activities
In the framework of this fellowship, I worked on a Technical Report (TR) that addresses critical gaps and challenges in the international standards landscape for digital content identification and binding mechanisms.
The absence of a common terminology across standardisation communities poses a major challenge. Different communities use inconsistent language when describing how content is connected with its metadata or other associated information. Whereas the C2PA initiative uses its own distinct terminology, other standardisation communities (e.g. W3C or OAIS) have different interpretations of what bindings mean. This terminological divergence leads to interoperability and mutual understanding barriers. The TR is establishing a comprehensive taxonomy that provides a neutral reference framework for multiple standardisation efforts, facilitating clearer communication across standardisation communities.
A gap the TR is addressing, is the limited comprehension of how binding mechanisms respond to content transformations. Digital content undergoes frequent alterations through compression, format conversion, and editing. Traditional identifier systems often fail when these changes occur, particularly when embedded metadata is stripped. The Working Group systematically analyses characteristics and limitations of different binding approaches, from cryptographic hashing to robust fingerprinting to watermarking techniques. This analysis will help stakeholders to make informed architectural decisions tailored to their specific requirements.
Moreover, the fellowship further contributes to positioning the recently published ISCC standard (ISO 24138:2024) within a broader global context. The TR serves as an educational resource, helping stakeholders understand how similarity-preserving identification methods complement established identification systems and address emerging needs in content provenance and authenticity verification, particularly relevant with current growth of AI-generated content.
Impact on SMEs (9th Open Call)
The TR will guide SMEs in understanding binding mechanisms: structural (metadata embedding), semantic (descriptive relationships), algorithmic (hashes, content-derived identifiers), and resolvable (URLs, DOIs).
In terms of applications, an Italian start-up, amlet.ai, adopted ISCC (one algorithmic binding approach examined in the TR) for their TDM registry. Also, Dutch liccium.com implements ISCC for decentralized content registration and rights management. Estonian valunode.com uses ISCC in their decentralised content management solutions. These implementations exemplify relevance across AI/TDM, rights management, and digital content workflows.
In terms of Impact, the TR clarifies how embedding, watermarking, fingerprinting, and cryptographic approaches differ in robustness and workflow requirements, helping SMEs make informed decisions and build expertise. Content-derived methods computing identifiers locally enable GDPR-compliant implementations without centralised tracking, supporting digital sovereignty.
Impact on society (9th Open Call)
I can see several societal impacts for this work, including:
Digital Trust and Information Integrity: The TR systematically documents capabilities and limitations of different content binding mechanisms and enables an informed selection of appropriate trust mechanisms, critical for democratic processes and media trust in the AI era.
Data Sovereignty and Privacy: The analysis of decentralised identification methods directly supports European digital sovereignty principles and GDPR compliance. By documenting alternatives to centralised tracking, the work enables implementations where rightsholders maintain control over digital assets while supporting privacy-by-design standards, addressing fundamental European values around data protection.
